Differential solar rotation depends on solar activity

SUNSPOTS have long been used as tracers to determine the rotation rate of the Sun. Scheiner 1 noted 250 years ago that low-latitude sunspots rotate more rapidly than those at high latitude (‘differential rotation’). In a recent review, Howard 2 stressed the limitations of the sunspot technique for determining rotation rate and differential rotation, as long-lived spots may have substantial individual motions, random and systematic, in both longitude and latitude. For example, leading and following spots of a sunspot group diverge early in the group's lifetime 3 ; rotation rates for isolated (‘unipolar’) spots and groups differ, with group rates depending on the shape of the group 4 . Another problem in using sunspots is the Wilson effect, a surface depression of the spot causing errors in longitude estimates away from the central meridian. Despite these difficulties, spots have been used previously to establish that there is no secular variation in rotation rate and differential rotation, nor any variations with phase of the solar cycle 5 , the only possible exception being during long-term disruptions in the solar activity cycle mechanism 6 . We present here evidence to suggest that the first of these conclusions is incorrect, and that differential rotation does depend on the amplitude of the solar cycle, at least for unipolar spots.