Based on satellite-derived global snow cover data on weekly time scales, the climatology and interannual variability of snow onset day-of-year, snowmelt day-of-year and number of snow-free days in a year are presented. Trends for snow onset day-of-year, snowmelt day-of-year and number of snow-free days in a year indicate that there has been an increase in number of snow-free days in recent decades. The relationship between snow cover and the Arctic oscillation (AO) index is examined on a hierarchy of time scales using lagged correlation and composite analysis. On weekly time scales, composite snow extent anomalies are maximum when AO leads snow cover by 1 week. These composite differences are maintained several weeks thereafter, particularly in the negative phase of the AO. Maps of composite snow cover anomalies when AO leads snow cover by 1 week delineate the spatial structure of these snow anomalies. On monthly time scales, lead–lag correlation between monthly snow cover and AO index indicates that the AO index during January, February and March is significantly correlated with snow cover in concurrent and subsequent spring months, particularly over Eurasia. Finally, on seasonal time scales, it is shown that winter season AO and winter/spring season snow cover are significantly correlated. Copyright © 2003 Royal Meteorological Society.