Auxiliary Material for Paper 2013GL058084A signal of persistent Atlantic multidecadal variabilityin Arctic sea iceMartin W. MilesUni Research,Bergen, Norway Bjerknes Centre for Climate Research,Bergen, NorwayInstitute for Arctic and Alpine Research, University of Colorado,Boulder, Colorado, USADmitry V. DivineNorwegian Polar Institute, Polar Environmental Centre,Troms¿, NorwayDepartment of Mathematics and Statistics, University of Troms¿,Troms¿, NorwayTore FurevikGeophysical Institute, University of Bergen,Bergen, NorwayBjerknes Centre for Climate Research,Bergen, NorwayEystein JansenDepartment of Earth Science, University of Bergen,NorwayBjerknes Centre for Climate Research,Bergen, NorwayMatthias MorosBaltic Sea Research Institute,Rostock, GermanyAstrid E. J. OgilvieInstitute for Arctic and Alpine Research, University of Colorado,Boulder, Colorado, USAMiles, M. W., D. V. Divine, T. Furevik, E. Jansen, M. Moros, and A. E. J. Ogilvie (2014), A signal of persistent Atlantic multidecadal variability in Arctic sea ice, Geophys. Res. Lett., 41, doi:10.1002/2013GL058084.IntroductionThe auxiliary material contains supporting information in the form of text and figures that provide details of the data and methods used in the analysis, with references. This material is provided as a single PDF file.1. 2013GL058084-txts01.pdfText S1. Descriptions of (1) sea ice time series used in the analysis, with Figure S1, (2) time-series analysis methods used in the analysis, and (3) Icelandic sea-ice proxy analysis, with Figures S2 and S3.Figure S1. Location map of historical and paleo proxy sea-ice time series analyzed here. White dots signify the approximate locations of the eight historical records, whereas brown dots signify the two sea-ice proxy records. Letters correspond to Table 1 and the panels in Fig. 1 in the main paper. Bathymetry and topography are shown in relief, and selected Arctic marginal seas are indicated. Figure S2. (a), Detrended MD99-2269-core sea-ice proxy (SIP) series (gray) and its multidecadal (50Ð120 year) component (blue). The wavelet bandpass-filtered component represents some 15% of the signal variance with respect to the variance of the resampled time series. (b) Sampling increment of the MD99-2269 sea-ice proxy series. The dark gray bars in (a) denote those parts of the record where the low-sampling increment precludes estimation of multidecadal variability.Figure S3. REDFIT estimates of the power spectral density function for the detrended MD99-2269 sediment core (0Ð6 kyr BP) series of sea-ice proxy (black solid lines); 95% and 90% Ôfalse alarmÕ levels (gray dashed-dotted and dotted lines, respectively) for the theoretical AR(1) spectrum calculated from the percentiles of the Monte Carlo ensemble.