The problem of solar wind data correlation for the observations onboard several spacecraft is very important for three points of view:
This problem is studied now in detail by systematic determination of ion flux and ion velocity data correlations by INTERBALL-1 (IKI group), WIND and IMP 8 (both - MIT group).
At the Fig. 1,2,3,4 one of examples of such comparing for the 6-hours interval on April 3, 1996 is presented.
Fig. 1 shows the positions of three spacecraft at GSE-coor dinates in the projection to the ecliptic plane. WIND was ahead of the Earth at about 75 Re along the Xse; INTERBALL and IMP 8 were separated by about 50 Re along Yse.
Fig. 2 shows the comparing of simultaneous ion flux measurements onboard three s/c with time resolution 1-1.5 min. Data for WIND and IMP 8 are shifted by the time of solar wind propagation to the INTERBALL-1 position. It can be seen the very good coincidence of these data including many fine features.
Fig. 3 shows the dependence of cross-correlation coefficients (INTERBALL-1 / WIND and INTERBALL-1 / IMP 8) on the time lag. For both comparings correlation coefficient is very high - up to 0.96 (!) - and its maximum is near zero lag. It means that in this event: a) the solar wind structures aren't changed significantly during about 20 min. (propagation time from WIND to INTERBALL-1); b) correlation length is more than 50 Re in the direction perpendicular the Xse.
Fig. 4 shows the similar comparing for solar wind velocities. Also the data from three spacecraft coincide very well.
This work is in progress now and it's possible to present as a preliminary result the hystogramm of cross-correlation coefficients for comparing of INTERBALL-1 and WIND ion flux data in 76 events (each event contents 6 hours of measurements - about 330 points) during April, 1996.
Such distribution is shown at Fig. 5. It can be seen, that in the 50% of events the cross-correlation coefficient is very high - more than 0.82. The average value of correlation coefficient for this period (one month) is 0.76. But there are a few events (about 20%) with low correlation (coefficient less than 0.6). The reasons of the low correlation events observed sometimes now is one of the targets of our study.