We have analysed the temporal characteristics of Arctic sea-ice extent and area in terms of its lagged correlation (an indicator for memory and variance) at monthly/seasonal timescales both in observations and a general circulation model (GCM) ensemble. We find that both observations and model generally match and exhibit a red noise spectrum in which significant memory is lost within 2–5 months. Beyond this time there is an increase in correlation, known as a re-emergence of memory, which is more ubiquitous in the model than observations. There are two distinct cycles of re-emergence in the model, one driven by sea-ice coverage–thickness coupling and the other by sea-ice coverage–SST coupling. The former shows up spanning 1 year as a summer-to-summer persistence, which is absent in observations. The source in the model is persistence of sea-ice thickness in the central Arctic and its influence on summer ice areal coverage. The second cycle of re-emergence results from correlation between pairs of months that share a similar mean sea-ice coverage, from months in the melt half of the season to months in the growth half of the seasonal cycle of sea-ice coverage, and is strongest during the late summer. The mechanism for this seasonal memory is via anomalous SSTs that persists at the location of the original sea-ice anomaly during the melt season and then alters the rate of sea-ice growth when the ice returns to the region. This behavior is also seen in observations. These patterns of memory re-emergence reflect on the seasonal cycle of the 1 month lagged correlation throughout the year, with low values in spring and autumn and high values during summer and winter. September sea-ice extent is significantly correlated with the previous August and July (and thus these months show a predictive skill of summer minimum), yet uncorrelated with months from November to June. Only limited predictive skill of September sea-ice coverage is gained from thickness anomalies during the winter in the model. We also show that memory re-emergence is enhanced by the sea-ice-albedo feedback.