IMAGE COALIGNMENT AND A FLOW MAP CALCULATION


The MDI images were accurately co-aligned. To remove motion of the entire group due to the solar rotation we use the large sunspot, A, as a reference object. We used a local correlation tracking method to make the required measurements. The FWHM of the Gaussian window (tracking window) was 6 x 6 arcsec. A flow map was calculated for each pair of successive images in the data set and is based upon a 20 min correlation interval.

By tracking the MDI intensity structures with this time interval we primarily measure displacements of pores and small spots and reduce contributions from the motions of shorter lived granules. Compared to the granulation, pores and sunspots are large and high contrast features, which makes them more easy to track.

Our error estimate is based on spatial and temporal consistency of the flow maps. Using a 44 hour intensity movie, we selected an extended area which showed a stable flow pattern. The error is defined as the minimum velocity value for which the velocity vector is guaranteed to be in the correct quadrant. Our estimation is that the solar noise is less than 30 m s-1. The final map was derived from the 44 hour average (136 flow maps) of the local cross correlation, which eliminated all short-term oscillations in the horizontal velocity field as well as reducing noise.


©BBSO/NJIT 2001