The Butterfly Diagram

The lower image in this figure is a butterfly diagram for the time period 1900–1993.

The construction of sunspot butterfly diagrams was first carried out by E.W. Maunder in 1904, and proceeds as follows: one begins by laying a coordinate grid on, for example, a solar white light or calcium image, with, as in the case of geographic coordinates on Earth, the rotation axis defining the North-South vector. The visible solar disk is then divided in latitudinal strips of constant projected area, and for each such strip the percentage of the area covered by sunspots and/or active regions is calculated and color coded. This defines a one-dimensional (vertical) array describing the average sunspot coverage at one time. By repeating this procedure at constant time intervals and stacking the arrays one besides the other, one obtains a two-dimensional image of average sunspot coverage as a function of heliospheric latitude (vertical axis) and time (horizontal axis). Tilted 90 degrees to the side, the diagram reminds one of a row of butterflies, thus the name butterfly diagram. Several features of these diagrams are noteworthy; the absence of sunspots at high latitudes (≥ 40 degrees) at any time during the cycle, and the equatorward drift of the sunspot distribution as the cycle proceeds from maximum to minimum are particularly striking here. Note how the latitudinal distribution of sunspots is never exactly the same, and how for certain cycles (for example cycle 20, 1965-1976) there exists a pronounced North-South asymmetry in the hemispheric distributions. Note also how, at solar minima, spots from each new cycle begin to appear at mid-latitudes while spots from the preceding cycle can still be seen near the equator, and how sunspots are almost never observed within a few degrees in latitude of the equator. Sunspot maximum (1991, 1980, 1969, ...) occurs about midway along each butterfly, when sunspot coverage is maximal at about 15 degrees latitude.

The plot on the upper right is a butterfly-like diagram running from the maximum of cycle 21 to one year of cycle 22 maximum. The latitudinal distribution of helmet streamers (blue +'s) illustrates the response of the corona to changes in the surface magnetic field, and can be seen to vary in phase with the sunspot distribution during that time interval. Note, however, that streamers are observed all the way to the poles at solar maximum, and that at any given time in the cycle streamers extend to higher latitudes than sunspots and active regions (see also slide 12). A butterfly-like diagram for the observed latitudes of coronal mass ejections (green 's; see also slide 13 and slide 14) nearly coincides with this diagram, with mass ejections occurring all the way to polar regions near the sunspot maximum phase of the cycle. Note that the absence of mass ejections from 1981 to 1984 is an artifact; the lack of mass ejection data from September 1980 to June 1984 is due to a breakdown of the SMM satellite; it was successfully repaired in space by the crew of the space shuttle Challenger in April 1984. A butterfly-like diagram for the observed latitudes of solar flares (not shown), on the other hand, coincides with the sunspot butterfly diagram. This further supports the notion that the larger-scale coronal manifestations of solar activity (such as coronal mass ejections) are not directly related to flares and other smaller-scale phenomenon confined to active region latitudes. Solar variability and solar activity must then be taken as the totality of processes operating within the Sun and their observed effects over the entire solar surface and at the Earth.

Written By P. Charbonneau and O.R. White–April 18, 1995