ISSN 1063-7729, Astronomy Reports, 2011, Vol. 55, No. 5, pp. 463–471. c  Pleiades Publishing, Ltd., 2011. Original Russian Text c  I.S. Knyazeva, N.G. Makarenko, M.A. Livshits, 2011, published in Astronomicheskii Zhurnal, 2011, Vol. 88, No. 5, pp. 503–512. Detection of New Emerging Magnetic Flux from the Topology of SOHO/MDI Magnetograms I. S. Knyazeva 1* , N. G. Makarenko 1** , and M. A. Livshits 2*** 1 Main Astronomical (Pulkovo) Observatory, Russian Academy of Sciences, Saint Petersburg, Russia 2 Institute of Terrestrial Magnetism, Ionosphere and Radio Wave Propagation, Troitsk, Moscow Region, Russia Abstract—A topological method for detecting the new emergence of magnetic flux using SOHO/MDI magnetograms of the full solar disk is proposed. This method uses the number of pixels in the image that can be distinguished from a specified value to within a predetermined threshold (the number of disconnected components). We study more than ten very powerful active regions (ARs) with very high flare activity and show that the number of disconnected components increases directly before the development of a series of M and X flares, or accompanies this process. This behaviour is evident not only when there is an explicit emergence of a new flux and a series of fast flares, such as in AR 9236 (November 2000), but also in groups with many non-stationary processes developing along a neutral line of the large-scale magnetic field. We also discuss the possibility of using the obtained results for flare prediction. DOI: 10.1134/S1063772911050040 1. INTRODUCTION The development of the techniques for solar magnetic-field measurements used on the SOHO and HINODE space observatories makes it possible to record the emergence of new magnetic fields. The huge amount of data that has been obtained with the SOHO MDI magnetograph with a temporal reso- lution of 96 min makes it possible in some cases to directly view the process of magnetic-flux emergence. This indicates the appearance of horizontal tubes of emerging new magnetic flux in active regions (ARs) (see, for example, [1]). There is reason to believe that the emergence of new magnetic flux plays an important role in the gen- eration of solar flares and coronal mass ejections [2, 3]. Flows in the form of magnetic flux tubes are gen- erated by the dynamo mechanism, and their dynam- ics can be investigated analytically in the thin-tube approximation [4] or numerically using MHD simu- lations [5]. Unfortunately, only large-scale structures representing a set of individual tubes can be moni- tored, and it is not possible to distinguish small-scale, subphotospheric flux structure [3, 6]. Therefore, it remains difficult to obtain a reasonable agreement be- tween numerical models of the subphotospheric mag- netic field and high-resolution observational data [7]. * E-mail: iknyazeva@gmail.com ** E-mail: ng-makar@mail.ru *** E-mail: maliv@mail.ru By tracking the evolution of ARs with a sequence of HINODE and SOHO/MDI image, the emergence of new magnetic flux can be detected directly. How- ever, even in this case, it remains unclear how to define the emergence at this initial stage, which is not visible by eye, and how to estimate the “inten- sity” and the “speed” of the flux, which is needed for comparison with the models [5]. Therefore, the early detection of new magnetic flux in the monitoring stage is considered as a separate problem, which can be approached mainly using modern techniques developed for digital image processing [8]. The re- cent paper [9] proposed the use of a Kalman filter and the SVM method for the automated detection of emerging flux in a sequence of magnetograms. A multifractal technique for identifying new flux was proposed in [10]. Singular points in the magnetogram are used as tracers of the new flux, obtained from estimates of the Holder exponents of the digital MDI magnetograms [11]. In this paper, we consider a topological approach to detecting new magnetic flux based on the analysis of connected structures (patterns) in the images [12– 15]. Its advantages are the possibility of detecting the flux emergence at early stages and tracking the evolution of the flux with time. We compare this evolution and flare activity for a number of solar active regions. 463