The contribution of the atmospheric circulation to a record minimum extent of open water in the polar Ross Sea (RS) region in the 2003 austral summer is examined. Two major findings are reached in this study. The first is that the origins of this anomaly are more complex than previously thought, with an anomalous atmospheric circulation contributing at least as much to the lack of open water as damming of sea ice by a large iceberg known as C-19. Only in the western RS, where C-19 lay, is damming found to restrict open water in the spring of 2002 (October-December), but even here the coldest spring in the last 15 years extended the sea-ice formation season. Elsewhere in the RS the divergent northward ice drift that normally occurs widely reversed to southward in early spring and was then followed by negligible ice motion. The most anomalous springtime ice drift occurred in the central and eastern RS rather than near C-19 and was mirrored in the weakest southerly winds on record in central areas. The unusual southward ice drift in early spring 2002 caused widespread convergence and compaction of the normally thin and undeformed first-year ice along and north of the central and eastern Ross Ice Shelf (RIS). Compacted ice with few leads would have been slow to melt in the warmest summer (January-February) months. Direct observations also indicate that sea surface temperatures (SSTs) rapidly fell to freezing in the central and western RS in February 2003 supporting new ice formation. All the available data indicate these were due to the cold spring and the extensive, compact ice cover in the late spring of 2002 and in January 2003 preventing most of the incoming solar radiation from reaching the ocean. The second finding is that sea ice in the summer of 2003 would have been very extensive even if C-19 had not occurred. This is based on comparisons with other years of extensive summer ice, notably one when no large iceberg occurred. Ice motion and atmospheric circulation patterns in this case resembled those in the spring of 2002 and the summer of 2003. Evidence that the anomalous atmospheric circulation in the RS in the spring of 2002 was typical of El Niño events is also discussed.