Fig. 1. Schematic map with 500 m contour intervals; the 10 m contours in the Dome C and Talos Dome area are derived from a Digital Elevation Model provided by Brisset and Rémy (1996).

Fig. 2. Comparison of snow accumulation from snow radar (L1 calibrated using the depth–age function at EPICA) versus snow accumulation derived from firn core analysis (A17β: Beta marker and A17t: tritium marker) within a 25 km radius of Dome C.

Fig. 3. Talos Dome snow accumulation map (kg m^− 2 yr^− 1) based on snow radar data (S2: 1835–2001 AD). Contour lines of surface topography have a 5 m interval. Snow radar profiles (dotted line), vector ice velocity (red arrow) and prevalent wind direction (blue arrow) are shown. The cyan line indicates the elevation and snow radar profile of Fig. 4.

Fig. 4. Talos Dome snow radar profile from GPR23 to GRP26 (position is shown in Fig. 3). Elevation, S1 snow radar elevation profile (numbers indicate the depth of the layer) and the 81 yr average of snow accumulation inferred from GPR are shown. Difference in accumulation inferred from GPR for three different periods are shown with error bars.

Fig. 5. Palaeo-accumulation (S3 less S2: 1635–1835 AD grey scale) and present accumulation (surface less S2: 1835–2001 AD red contour) maps based on snow radar data. Green, blue and black arrows indicate vector of ice velocity acceleration–deceleration and direction of rotation bearing. Contour lines of surface topography have a 5 m interval. Snow radar profiles (dotted line) and prevalent wind direction (cyan arrow) are shown.

Fig. 6. Dome C snow accumulation map (kg m^− 2 yr^− 1) based on snow radar data (L5: 1710–2000 AD) and snow accumulation values (in red) derived from firn core analysis (tritium/β marker: 1965–2000 AD). Contour lines of surface topography are placed at 2 m intervals. The dashed line indicates snow radar profiles. The cyan line indicates the elevation and snow radar profile of Fig. 7.

Fig. 7. Dome C snow radar profile from SW to NE along the major axis of Dome C (position is reported in Fig. 6). Elevation, L5 snow radar elevation profile (numbers indicate the depth of the layer) and the 290 yr average of snow accumulation profile are inferred from GPR. Differences of accumulation inferred from GPR for three different periods are shown with error bars.

Fig. 8. Temporal variability (5 data average) of D excess (‰) and sodium at EPICA EDC96 (from IGBP PAGES/World Data Center for Paleoclimatology Data Contribution Series #2005-048; [Röthlisberger et al., 2000] and [Stenni et al., 2001]; Stenni pers. com. 2005), D excess of old Dome C (Lorius et al., 1979) and EPICA EDC96 snow accumulation (1550–1998 AD) from both nssSO₄²⁻ volcanic spikes (Castellano et al., 2005) and stake farms (1996–2005 AD).

Fig. 9. Dome C snow palaeo-accumulation map based on snow radar data (L6 less L5: 1602–1739 AD). Red contour lines of accumulation map from present back to 1739 AD ( Fig. 7). Contour lines of surface topography are placed at 2 m intervals. The dashed line indicates snow radar profiles.

Position (2002 AD) of the stakes at Talos Dome, annual movement (velocity, direction) and acceleration–deceleration (from 1996–98 to 2002–05 AD)