Magnetic field observations made during 28 October to 1 November 2003, which included two fast interplanetary coronal mass ejections (ICMEs), allow a study of correlation lengths of magnetic field parameters for two types of interplanetary (IP) structures: ICMEs and ambient solar wind. Further, they permit the extension of such investigations to the magnetosheath and to a distance along the Sun‐Earth line (X) of about 400 RE. Data acquired by three spacecraft are examined: ACE, in orbit around the L1 point; Geotail, traveling eastward in the near‐Earth solar wind (at R ∼ 30 RE); and Wind, nominally in the distant geomagnetic tail (R ∼ −160 RE) but making repeated excursions into the magnetosheath/solar wind due to the flapping of the tail. Analyses are presented in both time and frequency domains. We find significant differences in the cross‐correlation/coherence properties of the ambient interplanetary magnetic field (IMF) and ICME parameters. For the ambient IMF, we find high coherence to be confined to low frequencies, consistent with other studies. In contrast, ICME magnetic field parameters remain generally coherent up to much higher frequencies. Scale lengths of ICME magnetic field parameters are in excess of 400 RE. High speeds of ∼1700 km s−1 are inferred from the plot of phase difference versus frequency, consistent with that obtained from plasma instruments. To strengthen these results and to extend them to include dependence on the distance perpendicular to the Sun‐Earth line (Y), we examine a 28‐day interval in year 2001 characterized by a sequence of 10 ICMEs and containing roughly equal ambient solar wind and ICME time intervals. ACE‐Wind X and Y separations were ∼220 and ∼250 RE, respectively. We find good coherence/correlation alternating with poor values. In particular, we find that in general ICME coherence/correlation lengths along Y are larger by a factor of 3–5 than those quoted in the literature for ambient solar wind parameters. Our findings are good news for the space weather effort, which depends crucially on predicting the arrival of large events, since they make possible the placement of upstream monitors to give a longer lead time than at L1.
