Determination of horizontal velocity fields at the sun's surface with highspatial and temporal resolution

We analyse the two algorithms which have been used in the past few years to determine the horizontal flow fields at the Sun's surface, namely the Loca l Correlation Tracking (LCT) of L. November and the Feature Tracking of L. Strous. Analysing the systematic errors introduced by LCT, we show that the se errors come from the averaging processes. More precisely, they arise fro m the interpolating step of the algorithm: granules' motions determine the flow on an irregular grid which is then interpolated to derive quantities s uch as horizontal divergence or vertical vorticity. Interpolation is theref ore a crucial step since mesoscale structures have mainly been studied thro ugh divergences and vorticities. We conclude that a reliable algorithm shou ld be based on the tracking of coherent structures, like granules, since th ey are representative of the fluid motion, and should contain an interpolat or which keeps track of the errors introduced either by location of the dat a (the shape of the irregular grid) or by the noise in the data.