Arctic Climate Spatio-temporal Modes of Variability as Sources of Predictability

The Arctic region has suffered a transformation in the past decades that will very likely continue in the future. Since the late 1970s, declining trends in pan-Arctic sea ice extent and volume, and increasing trends in air and sea surface temperature have been observed. However, the region has a large natural climate variability that can usually be mistaken for a long-term forced response. Furthermore, this natural Arctic variability has been linked to mid-latitude weather extremes in the northern hemisphere, both as a cause and a as response. Disentangling natural variability and forced response is of critical importance from a climate change perspective.

Using observational evidence and advanced statistical methods (K-mean clustering), we studied natural modes of variability in key Arctic climate variables (i.e. sea ice concentration, sea surface and near surface air temperature). The main objective was to identify spatio-temporal coherence in the series and lag-lead correlations between variables after detrending them.
Sea ice concentrations show three main modes of variability, as previously found. Both atmospheric near surface temperature and sea surface temperatures can individually affect sea ice and partly explain some regional patterns of natural variability on seasonal timescales. Pan-arctic sea ice extent shows a delayed response of 1-3 months to near surface air temperature and about a year to sea surface temperature variations.