DYNAMICS OF THE CHROMOSPHERIC NETWORK - MOBILITY, DISPERSAL, AND DIFFUSION-COEFFICIENTS

Understanding the physics behind the dispersal of photospheric magneti c flux is crucial to studies of magnetoconvection, dynamos, and stella r atmospheric activity. The rate of flux dispersal is often quantified by a diffusion coefficient, D. Published values of D differ by more t han a factor of 2, which is more than the uncertainties anew. We propo se that the discrepancies between the published values for D are the r esult of a correlation between the mobility and flux content of concen trations of magnetic flux. This conclusion is based on measurements of displacement velocities of Ca II K mottles using an uninterrupted 2 d ay sequence of filtergrams obtained at the South Pole near cycle minim um. We transform the Ca II K intensity to an equivalent magnetic flux density through a power-law relationship defined by a comparison with a nearly simultaneously observed magnetogram. One result is that, wher ever the network is clearly defined in the filtergrams, the displaceme nt vectors of the mottles are preferentially aligned with the network, suggesting that network-aligned motions are more important to held di spersal than deformation of the network pattern by cell evolution. The rms value of the inferred velocities, R = [\upsilon\(2)](1/2), decrea ses with increasing flux, phi, contained in the mottles, from R approx imate to 240 m s(-1) down to 140 -1 m s(-1). The value of R(phi) appea rs to be independent of the flux surrounding the concentration, to the extreme that it does not matter whether the concentration is in a pla ge or in the network. The determination of a proper effective diffusio n coefficient requires that the function R(phi) be weighted by the num ber density n(phi) of mottles that contain a total flux phi. We find t hat n(phi) decreases exponentially with phi and propose a model of con tinual random splitting and merging of concentrations of flux to expla in this dependence. Traditional methods used to measure D tend to be b iased toward the larger, more sluggish flux concentrations. Such metho ds neglect or underestimate the significant effects of the relatively large number of the more mobile, smaller concentrations. We argue that the effective diffusion coefficient for the dispersal of photospheric magnetic flux is similar to 600 km(2) s(-1).