Iridium 30 and 31 exchanged places in the constellation on September 19-22, 2002.
I would like to thank Walter Nissen for initially drawing my attention to this - see Status and Identity of Iridium 30 and Iridium 31 on SeeSat-L at http://www.satobs.org/seesat/Jul-2003/0151.html
Up to and including September 18, 2002, Iridium 30 was followed in the regular pattern by Iridium 31 :
Iridium 30 1 24949U 97051F 02261.14806094 -.00000231 00000-0 -89653-4 0 8872 2 24949 86.3995 17.9087 0002209 56.9878 303.1529 14.34213466262410 Iridium 31 1 24950U 97051G 02260.87532995 -.00000360 00000-0 -13554-3 0 8889 2 24950 86.3993 18.0069 0002271 58.9272 301.2147 14.34212368262382 IRIDIUM 30 1 24949U 97051F 02261.84574731 -.00000193 +00000-0 -76017-4 0 08944 2 24949 086.3998 017.6177 0002235 060.2566 299.8858 14.34213526262515 IRIDIUM 31 1 24950U 97051G 02261.85209058 -.00000211 +00000-0 -82404-4 0 08986 2 24950 086.3990 017.5985 0002292 063.4936 296.6506 14.34213018262526
A day later, Iridium 30 had been raised by 10km and Iridium 31 lowered by 10km. These changes appear to have taken place at around 03:00 UT on September 19, 2002:
IRIDIUM 30 1 24949U 97051F 02262.68297198 .00066488 00000-0 24949-1 0 8880 2 24949 86.4041 17.2727 0002372 77.9375 276.0562 14.31204631262633 IRIDIUM 31 1 24950U 97051G 02262.68931529 .00036298 00000-0 12095-1 0 8891 2 24950 86.4053 17.2485 0002117 83.7475 282.6644 14.37378306262644 IRIDIUM 30 1 24949U 97051F 02262.82399796 .00037339 00000-0 14069-1 0 8899 2 24949 86.4033 17.2127 0002324 77.5571 282.5854 14.31202862262651 IRIDIUM 31 1 24950U 97051G 02262.82733350 .00016341 00000-0 54583-2 0 8907 2 24950 86.4042 17.1891 0002128 86.9874 273.1520 14.37374441262668 IRIDIUM 30 1 24949U 97051F 02263.24349255 .00002886 00000-0 10863-2 0 8901 2 24949 86.4031 17.0390 0002388 81.1700 278.9537 14.31186517262716 IRIDIUM 31 1 24950U 97051G 02263.24502930 -.00000147 00000-0 -55923-4 0 8913 2 24950 86.4039 17.0140 0002177 89.5276 270.6120 14.37364776262724
On this basis, Iridium 31 would have overtaken Iridium 30 at around 14:30 UT on September 20, 2002.
IRIDIUM 30 1 24949U 97051F 02264.36216666 -.00003261 00000-0 -12451-2 0 8921 2 24949 86.4020 16.5744 0002307 79.9658 280.1786 14.31170241262874 IRIDIUM 31 1 24950U 97051G 02264.42850689 -.00002106 00000-0 -71311-3 0 8927 2 24950 86.4037 16.5201 0002285 89.2377 270.9082 14.37355916262897
Subsequent elsets show Iridium 31 preceding Iridium 30 in the regular pattern, the exchange having taken around 3 days:
IRIDIUM 30 1 24949U 97051F 02266.59952037 -.00107328 00000-0 -38984-1 0 8935 2 24949 86.4056 15.6428 0002675 71.8218 304.7160 14.34168198263199 IRIDIUM 31 1 24950U 97051G 02266.58661706 -.00070580 00000-0 -25474-1 0 8935 2 24950 86.4059 15.6161 0003051 92.4409 250.2627 14.34220528263209
Over the next few days, minor adjustments appear to have been made to get the satellites into precise new positions.
I have no idea of the reason for this changeover, and no official information has been forthcoming on the matter! Possibilities include a collision avoidance manouver or matching up failed cross-link antennae in adjacent orbital planes.